Friday, January 03, 2014

EHR Go-Lives Are Often Chaotic; One Area To Be Explored Is If This Go-Live Led To This Tragedy

EHR "go-lives" are particularly chaotic as staff adjusts to the new cybernetic governor of care.  Could the distractions have caused or contributed to the following tragedy?

Family wants to keep life support for girl brain dead after tonsil surgery
By Tom Watkins and Mayra Cuevas, CNN
updated 4:32 PM EST, Wed December 18, 2013

The mother of 13-year-old Jahi McMath, who was declared brain dead Thursday, three days after undergoing surgery to remove her tonsils, said Tuesday that the family should make the call.

... The surgery, which occurred December 9 [at Children's Hospital & Research Center in Oakland, California - ed.], initially appeared to have gone well, said Sandy Chatman, Jahi's grandmother who is herself a nurse and saw the girl in the recovery room. "She was alert and talking, and she was asking for a Popsicle because she said her throat hurt," Chatman said.

But Jahi was then moved to the intensive-care unit, and her relatives were denied access to the eighth-grader for 30 minutes; when they finally were allowed to see her, they knew something was wrong. "Upon entry, they saw that there was way too much blood," Chatman said.

"We kept asking, 'Is this normal?'" Sealey said. "Some nurses said, 'I don't know,' and some said, 'Yes.' There was a lot of uncertainty and a lack of urgency."

Sealey said that when Chatman noticed that her granddaughter's oxygen levels were dangerously low, she called for help.

But Jahi went into cardiac arrest. The medical staff performed chest compressions to revive her and gave her clotting medications, but nothing worked.

The girl's brain was severely injured by lack of oxygen.  I am not commenting on the reported dispute regarding removing life support.

I am commenting on my concern about a possible contributory role of a new EHR.

At my Jan. 2, 2014 post "Doctors' Dissatisfaction With EHRs May Be Early Warning of Deeper Quality Problems" ( I wrote of the distractions that physicians reported were caused by EHR systems such as:

... current EHR technology interferes with face-to-face discussions with patients; requires physicians to spend too much time performing clerical work; and degrades the accuracy of medical records by encouraging template-generated doctors' notes.

I had also noted nurse's concerns of "inevitable" patient injury due to EHR distractions, such as at:

  • and at other posts citing similar nursing complaints.

The cases cited above involve the "EPIC" EHR, but similar issues arise will most of the current EHR sellers' products, which are unregulated.  

For instance see the ECRI Institute's Deep Dive study of EHR risk at In a volunteer study (i.e., only a fraction of true incidents reported) of 36 ECRI PSO hospitals, 171 EHR-related "events" serious enough to cause harm were voluntarily reported in just 9 weeks.  8 of the "health IT events" were reported to have resulted in patient harm, and 3 were possibly related to patient deaths.  

From a press release from nurses at Affinity Medical Center (Ohio) on the nature of the problems:

... The programs are often counterintuitive, cumbersome to use, and sometimes simply malfunction. Nurses are finding that the technology is taking time away from patients and fundamentally changing the nature of nursing.” ... I’m concerned that the manner in which this technology is being implemented may pose serious disruptions in patient care.”

An open letter from nurses at an Ohio hospital, Affinity, on EHR "threats to patient safety."  Click to enlarge.

The EPIC EHR apparently had just recently "gone live" at Children's Hospital Oakland.

From  "Children's Oakland completes Phase 1 of $89 million electronic records system", Nov 20, 2013 (

Children's Hospital & Research Center Oakland has completed the first phase of an $89 million Epic Systems Corp. electronic health records system that links inpatient operations and an oncology/hematology clinic.

Other outpatient clinics are expected to come online in March or April, spokeswoman Melinda Krigel told the Business Times.

... The project's overall cost, $89 million, includes hardware, software and other implementation costs, including a separate SoftLab system that interfaces with the main electronic medical records system, Krigel said.

The official "go-live" date was Nov. 5.

See also the Children's Hospital Oakland Annual Report at  On page 21: 

... The Epic system will launch in November 2013 at Children’s inpatient facilities as well as in the Operating Room, the Emergency Department, the Day Hospital, and the Oncology/ Hematology Clinic.

I believe the possibility of clinicians being so distracted by computer data entry duties, and/or communications being impaired by the system's outputs, that this patient was left anoxic for a crucial period of time needs to be investigated.

In my view, in the differential diagnosis of clinical chaos in 2014, the chaos caused by healthcare IT needs to be a consideration.

My concerns may be shown unfounded in this case (I hope they are), and the injuries the result of other factors.  In consideration of the reported complaints from other organizations, however, not conducting an impartial investigating of a role of the new EHR in this tragedy would be, in my opinion, cavalier.

-- SS 

Jan. 5, 2014 Addendum:

From a court document cached here: , the following is written at p. 11-12:

... Originally the surgery was uneventful and MCMATH awoke from sedation in the recovery room speaking with hermother, Petitioner LATASHA WINKFIELD asking for a popsicle.  MCMATH was taken to the ICU and her mother was told to wait several minutes while they fixed her IV.

After being told several times that it would be just another 10 minutes, approximately 25-45 minutes after MCMATH was brought into the ICU, WINKFIELD went back and found her daughter sitting up in bed bleeding from her mouth.  It was evident that this had been transpiring for some time.  The nursing staff said “it was normal” and the mother stayed at the bedside as the bleeding grew increasingly worse.  The nurses gave WINKFIELD a cup/catch basin for MCMATH to bleed from her mouth into.  WINKFIELD asked for assistance and was told that this was normal and was given paper towels to clean the blood off herself and MCMATH.

The bleeding intensified to where copious amounts of blood were being expelled from MCMATH’s mouth and then nose.  MCMATH’s stepfather was also present and assisted in the attemps to stem/collect the blood.

Again, WINKFIELD asked for assistance, and a doctor, and was only given a bigger container to collect the blood and, later, a suction device to suction the increasing volume of blood.  The stepfather continued to suction while the mother went and got her mother, a nurse, to take over for her.  The grandmother saw what was happening and made multiple requests, and then a loud demand, for a doctor.

MCMATH shortly thereafter suffered a heart attack and fell into a comatose state.  She later was pronounced “brain dead”… 

"Heart attack" (i.e., primary myocardial infarction) in a 13-year-old sounds far less likely than exsanguination to the point of hypovolemic shock, severe hypotension, and cardiac arrest.  That such events transpired in an ICU, with family present and calling for help, suggests there were major clinician distractions of some sort at play.

A reader wrote me wondering if a new CPOE component could have caused delays in evaluation and treatment. When someone is dying, you simply cannot waste time 'clicking away', they wrote.

A reader also wrote me wondering if an EHR crash occurred at the time this patient was left with family to exsanguinate, causing clinical chaos.

That is a particularly interesting thought.  See the multiple posts at

In my opinion, these issues require investigation.

-- SS


Anonymous said...

Indeed, there are other unexpected premature deaths and serious injuries that have made the newswires that occurred within 4-6 weeks subsequent to a "go live" of a hospital wide EHR system with CPOE and CDS devices.

What happens during these acutely vulnerable periods?

In short, there is hospital wide mayhem as the prior systems and workflows of care suddenly disappear, replaced by, as you state, a radically different system using cumbersome, counterintuitive, time consuming, and distracting devices.

Patients are neglected and care is delayed during these times, not to mention overdoses, mis identifications, duplicate medications, etc.

For instance, Han et al reported a 2.4x increase in baby deaths (Pediatrics 2005) after the EHR in a Pittsburgh children's hospital was activated.

The FDA ought to be investigating these device associated deaths and injuries, especially that of Jahi in Oakland.

Anonymous said...

This is a really interesting speculation, Scott. Neither of us can know, but I would not be surprised. I guess it's something to warn our family and friends to ask about if selecting a time for an elective surgery.

Steve Lucas said...

I was saddened to just read of a three year olds soon to be death as the result of a dentist performing three root canals under sedation. (Would a doctor explain why a three year old is having root canals?) The family is now faced with the agony of watching this child slip away.

The staff was found not to be monitoring the child’s vitals and her oxygen levels fell. One does not have to wonder if the reliance on technology, and the demands of other technology, played a part in this tragedy.

This beautiful baby girl will be gone forever, and no computer will ever bring her back.

Steve Lucas

Anonymous said...

Go lives are risky for patients. Han et al described an unexpected increase in baby mortality after a go live ca 2003 at a children's hospital (see below).

Since then, there have been tragic cases of neglect during the 3 months following go lives, several of which made the news, although no one wanted to connect the dots between the EHR go live and the catastrophes.

In no particular order, there is a death in a stairwell of a patient, a death by ice on a hospital roof of a patient, deaths by heparin overdose of several patients, to mention a few, all occurring in the peri go-live period.

Surveillance for the adverse events that occur during the peri go-live period should be mandatory. I think that reporting and analyzing the deaths and injuries of patients during the first 6 months after a go-live to be appropriate.

InformaticsMD said...

Anonymous said...

I guess it's something to warn our family and friends to ask about if selecting a time for an elective surgery.

I personally would not want to undergo any procedure or even be in the hospital within at least 3 months of EHR rollout, and I was a CMIO once. You can take that for what it's worth.

-- SS

InformaticsMD said...

Steve Lucas wrires:

This beautiful baby girl will be gone forever, and no computer will ever bring her back.

The hyper-enthusiasts appear to believe the EHR harms are necessary sacrifices for the benefit of the collective.

That the harms are "anecdotal" and just "bumps in the road" - two terms EHR leaders at OHSU and Cleveland Clinic used to my challenges - are examples of an attitude that I believe is counter to what I learned about medical ethics at Boston University School of Medicine and Yale School of Medicine.

-- SS

Anonymous said...

The glossing over of these mishaps as mere "bumps in the road" is also antithetical to the concept of the sanctity of the individual as enumerated in our Nation's founding documents. Blind ideology existing in a moral vacuum.

Anonymous said...

I'm an implementer at Epic. We try very hard to prevent these things from happening, but nobody seems to care on the project leadership side (both hospital and Epic). There are not enough developers or development done to fix these risks, and it is my professional opinion that a large fraction (especially fast-paced ones) are mismanaged and understaffed. A lot of the hospital IT teams I worked with do not seem to care or know how to do their jobs. Epic sets a lower criteria for certification for their customer teams, and the training (even the expensive onsite training) is not very good. I've attended a lot of these classes. They do not prepare you for implementing the clinical applications as well as they should. I'm going to start speaking out at the risk of losing my job and being sued (Carl Dvorak regularly admonishes us to not post anything on the blogs), because I think hospitals and Epic (and other EMRs, possibly) are sacrificing lives for the sake of increasing revenue. It's immoral. At one Epic Systems hospital, I watched a child suffer for an hour because Epic implemented a new workflow that was hard to follow (even with the "tip sheet" provided) and nobody knew how to get the system to move forward and dispense the drug. The kid had a spinal tap. It was awful. I complain to my superiors and they (as far as I can tell) ignore these threats. The turnover is really high at Epic and it's not just whiny kids who are burnt out. I worry about my hospital's patients daily and I think I'm going to leave soon.